Morgan Stanley: AI Debt on Track to Hit $570B in 2026
Key insights
- Morgan Stanley projects AI-linked bond issuance will reach nearly $570 billion in 2026, with $236 billion already sold by May 31.
- Hyperscaler capex now runs close to 100% of operating cash flows versus a 10-year average of 40%, forcing companies into bond markets.
- AI-linked debt surpassed US banks as the largest investment-grade market segment by October 2025, changing passive bond fund composition.
Why this matters
AI practitioners and technical leaders now operate inside a capital structure where data center infrastructure is the largest single segment of the investment-grade bond market, meaning credit market conditions directly affect hyperscaler build schedules and cloud capacity availability. For founders and operators relying on hyperscaler commitments, Amazon's $364 billion cloud backlog and $200 billion in planned 2026 capital spending are funded partly by new debt, yet S&P has warned that Amazon will likely post negative free operating cash flow over the next two years, introducing refinancing risk that could constrain capacity if AI revenues disappoint. The passive fund mechanism makes this exposure systemic: $4.8 trillion in target-date fund assets track bond indexes that must proportionally absorb every new eligible AI bond, so a sector-specific credit event would propagate into retirement accounts without any individual investor choosing the exposure.
Summary
Morgan Stanley forecasts nearly $570B in AI-linked bonds for 2026, with $236B already placed by May 31, four times the 2025 pace.
Hyperscaler capex now runs near 100% of operating cash flows versus a 40% decade average, so bond markets fill the gap free cash flow can no longer cover.
Essentially: Amazon, Alphabet, Meta, Microsoft, and Oracle are the primary issuers driving this supply wave.
- Amazon set the record Canadian maple bond (C$14 billion) and the largest euro corporate bond ever (14.5 billion euros).
- AI debt topped US banks as the biggest JPMorgan U.S. Liquid index sector by October 2025.
- $4.8 trillion in target-date fund assets absorb AI bonds automatically through index tracking.
S&P warns Amazon faces likely negative free operating cash flow for two years while May 2026 inflation sits at 4.2%, the highest since April 2023.
Potential risks and opportunities
Risks
- Amazon faces an S&P-flagged likelihood of negative free operating cash flow for the next two years while carrying $364 billion in cloud backlog and $200 billion in planned 2026 capex, making its credit spread vulnerable if AI revenue growth stalls before refinancing cycles close.
- Super Micro Computer's 17% single-session share drop on its June 10 $7 billion equity raise signals equity markets are repricing AI hardware leverage risk, and issuers without investment-grade ratings could find both bond and equity financing windows closing simultaneously.
- With AI-linked debt now the dominant sector in the JPMorgan U.S. Liquid index, a sector-specific shock would force passive rebalancing that amplifies any selloff into the $4.8 trillion target-date fund ecosystem, affecting retirement savers who never chose the exposure.
Opportunities
- Fixed-income teams with AI infrastructure credit expertise, such as M&G Investments which tracked the $1.2 trillion outstanding figure, can differentiate by pricing covenant and depreciation risk across hyperscaler bonds before spreads fully reflect the leverage buildup.
- Mid-tier data center operators with clean balance sheets have a demonstrated window: Hut 8 closed a $4.25 billion bond and Keel Infrastructure a $458 million convertible deal on June 10 and June 9 respectively, confirming market appetite before May 2026 inflation at 4.2% tightens conditions.
- Non-USD credit desks in Canada and Europe gain pricing leverage as Amazon sets market benchmarks with the record C$14 billion Canadian maple bond and the largest euro corporate bond ever at 14.5 billion euros, establishing deal-structure norms for subsequent AI infrastructure issuers.
What we don't know yet
- What specific AI revenue thresholds would trigger the credit spread widening Morgan Stanley warns about: no monitoring criteria or stress-test parameters were disclosed in the forecast.
- Whether semiconductor companies' new amortizing bond structures are calibrated to actual AI hardware depreciation cycles: the article notes the structural shift but provides no projected repayment timelines against hardware refresh rates.
- OpenAI's planned 10-gigawatt Ohio campus is described as costing at least $500 billion with lease talks described as advanced, but no bond structure, co-investors, or financing timeline has been publicly disclosed.
Originally reported by techtimes.com
Read the original article →Original headline: Morgan Stanley: Global AI-Linked Debt on Track to Nearly Double to $570 Billion in 2026 — AI Bonds Now Largest Investment-Grade Sector